Li-ion Battery Modeling And Simulation With Equivalent Circuit-Simplified Electrochemical Model

Lithium-ion batteries have gradually gained more and more applications in the field of electric vehicles recently,owing to their high energy density,high power density,long life-cycle and less memory effect.However,lithium-ion batteries are still facing a number of challenges in practice,such as fierce ambient temperature variation during operation,the overcharging or over-discharging potentials,the accelerated degradation due to high-rate charging etc.Therefore,to maintain the battery system operate in an efficient,secure and prolong lifespan,it is necessary to develop an advanced battery management system(BMS).As one of the core technologies for BMS,the battery model accuracy and its real-time performance has been ranked as the key issue and research priority.This paper analyzes the advantages and disadvantages of existing equivalent circuit model(ECM)and electrochemical model.Based on these two models,we conducted the in-depth analysis and research in the time and frequency domain.In this thesis,the modeling work covers the battery’s external characteristics and internal electrochemical mechanisim,and can be expanded as the following five aspects:(I)Parameter identification of ECM and influencing factors analysis.Firstly,the model is analyzed in the discretization form,including the model error analysis,stability analysis,perturbation analysis of measurement uncertainty and parameter sensitivity analysis.Through the above analysis,it can enhance the in-depth understanding of the model parameter identification problem and can evaluate the influence of the sampling frequency and current/voltage sensor measurement precision on the model parameter identification accuracy.Then,the temperature effect,current/ voltage sampling precision and sampling time effects on the parameter identification accuracy are analyzed.Finally,to meet with the specific precisions for parameter identification and SOC estimation,the current/voltage sensor accuracy and the optimal sampling time are recommended.(II)State of charge estimation and influencing factors analysis.In this section,the model parameters and SOC are jointly estimated by the recursive least square method and the adaptive extended kalman filter.This hybrid algorithm can update the model parameters online,which will improve the model adaptability in different scenarios.Finally,the simulation is conducted to verify the SOC estimation precision by considering the influencing factors,such as a)model order,b)initial SOC setting value,c)covariance of process noise and measurement noise,d)EKF & AEKF algorithm,e)current sensor accuracy,f)voltage sensor accuracy and g)sampling time.(III)Simplified electrochemical model with modified boundary conditions and its transfer function derivation.The pseudo two dimensional electrochemical model(EM)can describe the electrochemical process of the battery,which will enrich our in-depth understanding.The major disadvantage of the EM lays on its heavy computation burden,which severely limits its usage in real-time control problems.In this section,the electrochemical model is simplified by modifying the boundary conditions of the electrolyte concentration diffusion equation.By this way,the electrochemical model will be greatly reduced,and we can achieve the transfer function results for the lithium ion concentration,electrical potential with respect to the current density in electrolyte phase.Furthermore,the Pade approximation method is used to give a numeratordenominator type of transfer function for the simplified electrochemical impedance model,which reduces the computation effort greatly,and makes it more suitable for real-time control application.Finally,the accuracy comparison of the simplified and traditional electrochemical model are carried out.The simulation reveals the validity of the simplified electrochemical model with high level.(IV)Impedance analysis and parameter sensitivity analysis for simplified electrochemical model.In this section,it mainly discusses the short-term dynamic impedance model and the full electrochemical impedance model.About the short-term dynamic impedance model,the influencing factors,such as the equivalent conductivity k_eff,the electrolyte concentration diffusion coefficient De,and the exchange current density i0 are analyzed.About the full electrochemical impedance model,the influences of the negative phase diffusion coefficient Ds_neg and the positive solid diffusion coefficient Ds_pos on the impedance spectra are investigated.The parameter sensitivity analysis of the simplified electrochemical impedance model provides basic knowledge for the parameter identification in time domain and the impedance spectrum estimation in frequency domain.(V)Realtime verification of the simplified electrochemical model(SEM)and establish the parametric relationship between SEM and ECM.Currently,the two main battery models(equivalent circuit model & electrochemical model)are independent from each other,and each has its own advantages and disadvantages.In this paper,the proposed simplified electrochemical model(SEM)has achieve the optimal balance between EM and ECM.Then,the SEM is verified upon the dspace platform with DS1401/1512/1517 controller,and the experimental results prove that the SEM has high accuracy and great realtime performance.The establishment of the parametric relationship between the SEM and ECM will promote their scientific relationship.For example,the electrochemical model parameters could be identified or calculated using the ECM parameters.